Wording Matters: The Effect of Linguistic Characteristics and Political Ideology on Resharing of COVID-19 Vaccine Tweets

• Published in: ACM transactions on computer-human interaction 2024-09, Vol.31 (4), p.1-23, Article 44
• Main Authors: Borghouts, Judith, Huang, Yicong, Hopfer, Suellen, Li, Chen, Mark, Gloria
• Format: Article
• Language: English
• Subjects: Applied computing; Collaborative and social computing; Human computer interaction (HCI); Human-centered computing; Law, social and behavioral sciences
• ISSN: 1073-0516; 1557-7325
• DOI: 10.1145/3637876

Social media platforms are frequently used to share information and opinions around vaccinations. The more often a message is reshared, the wider the reach of the message and potential influence it may have on shaping people’s opinions to get vaccinated or not. We used a negative binomial regression to investigate whether a message’s linguistic characteristics (degree of concreteness, emotional arousal, and sentiment) and user characteristics (political ideology and number of followers) may influence users’ decisions to reshare tweets related to the COVID-19 vaccine. We analyzed US English-language tweets related to the COVID-19 vaccine between May 2020 and October 2021 (N = 236,054). Tweets with positive and high-arousal words were more often retweeted than negative, low-arousal tweets. Tweets with abstract words were more often retweeted than tweets with concrete words. In addition, while Liberal users were more likely to have tweets with a positive sentiment reshared. Conservative users were more likely to have tweets with a negative sentiment reshared. Our results can inform public health messaging on how to best phrase vaccine information to impact engagement and information resharing, and potentially persuade a wider set of people to get vaccinated.